Network
Three regions, and the live list that proves it
Egress runs on machines in our own AWS account: us-east, us-west and europe. That is the whole fleet. This page says so plainly because a region list is the first thing anyone checks and the most expensive thing to be wrong about.
What we do not offer
There are no Asia-Pacific, Middle East or South America nodes, and no residential or mobile addresses. An earlier version of this site listed all five regions and three address types. If your work needs an exit in a region we do not run, we are not the right supplier today — that is a cheaper thing to learn here than after you integrate.
The regions
Each region is an egress node with a stable Elastic IP, provisioned from this repository and deployed by script. Requests you send to the API with a given region come out of that region's node.
us-east
AWS us-east-1. The default choice for North American targets.
us-west
AWS us-west-2. West-coast targets, and a second US exit when one is being rate-limited.
europe
AWS eu-west-1. European targets, and content that varies by continent.
The list you should actually trust
This page is prose and prose goes stale. GET /v2/proxy/regions is the authoritative list, and it is health-gated: a region with no healthy endpoint behind it is not advertised at all.
Failover is real and it is reported rather than hidden. If every endpoint in the region you asked for fails mid-request, the fetch is retried from another region and the response comes back with region_used set to the region that actually served it. If the exit country is part of what you are measuring, check that field rather than assuming you got what you asked for.
If you are building anything that depends on a particular exit, read that endpoint at start-up rather than hardcoding a region list from a marketing page.
Why the fleet is small on purpose
The previous version of this service routed through 33 third-party serverless endpoints whose source code we did not have. When they started failing, none of them could be debugged or fixed, and a re-probe found that zero of the 33 still worked.
Everything is now on infrastructure we own, with the node source in the same repository as the API. Three regions we can reason about beat thirty-three we cannot.
Questions
- Can I request a region you do not have?
- Yes — email support@roamingproxy.com. Adding a region is a scripted provision and deploy, so it is a question of demand rather than of engineering effort. We will tell you honestly whether and when.
- What happens if I ask for a region that is unavailable?
- A region with no healthy endpoint is not advertised at all, so it is not selectable. If the region you named is healthy when you call but all of its endpoints then fail, the request fails over to another region rather than failing outright — and the response tells you, in region_used. Read that field if the exit country matters to what you are measuring; it is the region that actually served the fetch, not the one you asked for.
Get a key
Create an account and mint an API key in the dashboard. The full endpoint reference — request shapes, parameters and error codes — is published at https://api.roamingproxy.com/v2/docs.